Waste grape skins thermal dehydration: potential release of colour, phenolic and aroma compounds into wine

Abstract

Exploitation of grape waste material is scarce. One of the main issues to deal with is its high moisture content, as it causes spoilage and degradation of valuable compounds. In order to assess this limitation, four different Vitis vinifera waste grape skins from the juice industry were dehydrated at 60, 90 and 100°C. Characterisation of dehydrated waste grape skins (DWGS) was firstly done in wine model solution to evaluate the release of colour, phenolic and volatile compounds. Colour of DWGS-60°C solutions was similar to the control one (freeze-drying) in terms of red and yellow components regardless of the variety. Samples of Garnacha Tintorera, Bobal and AMIX dehydrated at 60°C showed the highest content of low molecular weight phenolic compounds. Bobal DWGS-60°C released the highest concentration of volatiles. An assay by direct addition of Bobal DWGS-60°C into a white wine demonstrated the potential of these residues for producing rosé wines.

El aprovechamiento de residuos industriales de hollejos de uva es limitado. Una de las razones es el alto contenido de humedad, causante del deterioro microbiológico y de la pérdida de compuestos. Por esta razón, cuatro variedades de hollejos de uva procedentes de la industria de zumos fueron deshidratadas a 60, 90 y 100°C. Los hollejos deshidratados se caracterizaron primeramente en vino sintético, evaluándose la liberación de color, compuestos fenólicos y volátiles. Independientemente de la variedad, el color aportado por los hollejos deshidratados a 60°C fue similar al control (liofilización) en sus componentes roja y amarilla. Garnacha Tintorera, Bobal y AMIX deshidratadas a 60°C proporcionaron el mayor contenido en compuestos fenólicos de bajo peso molecular, mientras que Bobal–60° C lo fue para los volátiles. La adición directa de hollejos de Bobal–60°C a un vino blanco puso de manifiesto el importante potencial de estos residuos para producir vinos rosados.

Keywords:

• dehydration
• waste grape skins
• wine
• phenolic compounds
• aroma
• colour

Palabras clave:

• deshidratación
• hollejos
• vino
• compuestos fenólicos
• aroma
• color

Introduction

Grapes are one of the world's largest fruit crops with an annual production of nearly 67 million tons, figuring as the sixth largest commodity in Europe (FAOSTAT, 2011). Castilla-La Mancha region in Spain is considered one of the highest producing regions of grapes in the world with 2,905,591 metric tons (MT) (MARM, 2009) where almost 75% production is processed into wine and its balance into juice. During the grape harvest season, regional manufacturers generate approximately 523,000 MT of grape marc (a mixture of grape skins, seeds and stalks) as a residue. In the present time, part of this industrial waste is re-fermented and distilled to produce different spirits or processed into fertilizers (Arvanitoyannis, Ladas, & Mavromatis, 2006; Baran, Çayci, Kütük, & Hartmann, 2001; Cortés, Salgado, Rodríguez, & Domínguez, 2010; Ferrer et al., 2000; Ruggieri et al., 2009). However, disposing and incinerating substantial amounts of grape marc has become an important economic issue for producers as well as a significant environmental problem.

Grape skins are known as the major source of phenolic and aroma compounds when related to wine and juice quality (Canals, Llaudy, Valls, Canals, & Zamora, 2005; Gómez, Martínez, & Laencina, 1994; Kontoudakis et al., 2011). Although different grape processing operations aim to extract as much of these substances as possible, grape skins are not completely exhausted. In fact, waste grape skins (WGS) have a considerable amount of bioactive remaining substances, e.g. antioxidants, pigments, vitamins and aroma (Kammerer, Claus, Carle, & Schieber, 2004; Lafka, Sinanoglou, & Lazos, 2007; Ruberto, Renda, Amico, & Tringali, 2008). Grape waste from the juice industry is of particular interest because of its shorter extraction times used during processing and for receiving much less attention in comparison with those from the wine industry.

Given the current industry and consumer demand for natural ingredients and forthcoming sustainability regulations there is a growing interest in recycling and exploiting WGS (Bustamante et al., 2008; Cheng, Bekhit, Sedcole, & Hamid, 2010; Henry, Pauly, & Moser, 2001; Rózek, García-Pérez, López, Güell, & Ferrando, 2010). Recent works (Amendola, De Faveri, & Spigno, 2010; Corrales, Toepfl, Butz, Knorr, & Tauscher, 2008; Kammerer et al., 2004; Lafka et al., 2007; Spigno, Tramelli, & De Faveri, 2007) have focused on the characterisation and exhaustive extraction of winery wastes, evaluating different extraction methods, solvents and extraction conditions (temperature, solvent-to-solid ratio…) in terms of the phenolic composition. So far the applications and/or dissemination of these findings remain limited throughout the food industry as a big quantity of grape marc is still unexploited.

Pre-processing steps are needed to avoid spoilage as grape marc is susceptible to oxidation reactions as well as to microbial spoilage because of its high moisture content (60%). The stability and quality of stored grape marc for distillation has been evaluated in terms of different plastic and concrete containers (Cortés et al., 2010), because spoilage is responsible for off-flavours in the final product (Da Porto, 2002). Dehydration is an adequate alternative to prevent deterioration providing simultaneously mass and volume reduction. It also decreases other chemical degradation reactions and improves handling and storage (Singh & Heldman, 2009). Thermal dehydration processes are widely used (Ratti, 2001) involving physical, structural and chemical changes that affect important substances such as volatiles and polyphenols. Freeze-drying processing is an alternative method for some value-added foods because of the low damage caused during water removal by sublimation. This dehydration method has been employed as a reference when evaluating quality of dried fruits and vegetables (Larrauri, Rupérez, & Saura-Calixto, 1997; Ratti, 2001). Although freeze-drying is a prominent dehydration method, its high operating costs represent a limitation for its diffusion within the industry. Alternatively, some authors have found that low temperature thermal dehydration may preserve some properties and substances of interest nearly as well as freeze-drying (Larrauri et al., 1997).

The aim of this article was to evaluate how dehydration temperatures affect the composition of juice industry waste grape skins. Extraction of colour, phenolic and aroma compounds was done by direct addition of DWGS into wine model solutions in order to evaluate a potential application for the wine industry. Finally, the production of rosé wine was assayed by direct addition of dehydrated waste grape skins into white wine.

Materials and methods

Waste grape skins

Grape marcs from Vitis vinifera red varieties Garnacha Tintorera, Bobal and Cabernet Sauvignon, were obtained during the 2008 harvest season from a juice concentrate factory in Castilla-La Mancha (Julian Soler, Cuenca, Spain). Airen white variety was mixed with red variety (30–45%), namely AMIX and was sampled due to its high production level within the region. Inside the factory, each grape variety was crushed, destemmed, macerated for 3 days and mechanically pressed to obtain juice. Grape marc samples of 50 kg were collected in plastic bags immediately after pressing then frozen at −20°C until dehydration.

Chemicals and standards

Gallic acid, (+)-catechin, caffeic acid, ferulic acid, p-coumaric acid, vanillic acid, syringic acid, cis and trans-resveratrol, (−)-epicatechin and quercetin dihydrate from Sigma-Aldrich (Steinheim, Germany) were used as standards for low molecular weight phenolic compound analysis. Malvidin-3-glucoside (Mv-3-G) standard from Extrasynthése (Geneay, France) was used for anthocyanin quantification. D-limonene, linalool, linalyl acetate, α-terpineol, citronellol, nerol, geraniol, geranyl acetone, nerolidol, farnesol, trans-2-hexenal, 1-hexanol, trans -2-hexen-1-ol, α-ionone, β-ionone, 2-phenylethanol, isoamyl acetate supplied by Sigma-Aldrich (Steinheim, Germany) and β-damascenone supplied by Firmenich (Geneva, Switzerland) were used as calibration standards in model wine solution (12% ethanol, pH  =  3.6, 5 g L⁻¹ tartaric acid) for volatile analysis. HPLC-grade acetonitrile from Panreac (Barcelona, Spain).

Dehydration methods

Waste grape skins were placed in beds of 2.5 cm thickness and dehydrated in a conventional oven (Selecta, Barcelona, Spain) at 60, 90 and 100°C until constant moisture. For freeze-drying, frozen samples at −20°C were dehydrated in Telstar (Barcelona, Spain) LyoAlfa 6 equipment during 48 h. Heating plate temperature was 20°C, vacuum pressure 4 × 10⁻² mBa and condenser temperature −50°C. Freeze-dried samples were used as control treatment in order to evaluate the effect of dehydration temperature on composition of samples. For moisture content determinations, 5 g of waste grape skin samples were weighed with a precision balance (0.001 g) on a well dried Petri plate. The plate was then placed into an oven at 103°C±2°C until constant weight.

Direct addition of dehydrated waste grape skins into hydroalcoholic solution

Seeds and stalks were removed from dehydrated waste grape skins (DWGS) by means of sieving with a 3 mm mesh. Residual stems and seeds were removed by hand. Samples were then ground in a cutting mill MS 100 (Retsch, GmbH & Co. KG, Denmark) and sieved to a particle size of 1 mm. Five g L⁻¹ of DWGS were directly added to a model wine solution made with 5 g L⁻¹ tartaric acid, 12% ethanol and adjusted to pH = 3.6 with sodium hydroxide. DWGS-solutions were macerated at 18°C during 24 h in closed amber crystal bottles. Such extraction conditions were chosen for being potentially affordable in future procedures at winery industrial level. After maceration, the solutions were filtered with a strainer to remove DWGS and prepared for the various analyses. These conditions were also used for the assay of rosé wine using a commercial white wine of Airén variety (12% ethanol, pH = 3.6).

Chemical composition of DWGS-solutions

Colour and total phenolics by UV-vis spectrophotometry

Colour and total polyphenol index (TPI) have been measured in a Lambda 25 UV-Vis spectrophotometer (Perkin Elmer, Norwalk, USA) using quartz cells. Conventional wine parameters of colour intensity and shade were determined according to Glories (1984) by measuring the absorbance of DWGS-solutions at 420, 520 and 620 nm. Colour intensity was expressed by the sum of absorbance values while shade was expressed as a ratio of A420/A520. TPI was determined according to Singleton and Rossi (1965) by measuring the absorbance of DWGS-solutions at 750 nm. TPI results were expressed as mg of gallic acid equivalents according to a calibration curve prepared with the pure standard.

Low molecular weight phenolic compounds and anthocyanin determination by HPLC-DAD

The DWGS-solutions were filtered through a 0.45 μm PVDF Durapore filter (Millipore, Bedford, MA) and injected into an Agilent 1100 HPLC chromatograph (Palo Alto, USA) equipped with a Phenomenex Synergi 4μ Hydro-RP (4 μm particle size, 80Å pore size, 150 × 2.0 mm) (Torrance, USA) at 25°C following Cozzolino et al. (2004). Used solvents were; (A) 1% acetonitrile, 1.5% phosphoric acid in water and (B) 20% solvent A, 80% acetonitrile for gradient elution at a constant flow rate of 0.4 mL/min: 0 min (14.5% solvent B), 18 min (27.5% solvent B), 20 min (27.5% solvent B), 21 min (50.5% solvent B), 22 min (50.5% solvent B), 26 min (100% solvent B), and 28 min (100% solvent B). The injection volume was 20 μl. Compound detection was carried out through a diode array detector by comparison with the respective UV-vis spectra and retention time of pure standards. Quantification was done according to the maximum wavelength absorbance of each standard as follows; Gallic acid, (+)-catechin, vanillic acid, syringic acid and (−)- epicatechin were quantified at 280 nm, ferulic acid and caffeic acid were quantified at 324 nm, (trans)-resveratrol and p-coumaric acid were quantified at 308 nm, malvidin-3-glucoside was quantified at 520 nm. delphinidin-3-glucoside, cyanidin-3-glucoside, petunidin-3-glucoside, peonidin-3-glucoside together with their respective acylated derivatives were identified according to the elution order reported in the literature (Alcalde-Eon, Escribano-Bailón, Santos-Buelga, & Rivas-Gonzalo, 2006) and quantified as malvidin-3-glucoside equivalents. Quantification was based on 5-point calibration curves of their respective standards (R ² > 0.99) performed in a model wine solution previously described.

Sampling conditions for stir bar sorptive extraction

Stir bar sorptive extraction (SBSE) was used for analysing free and bound volatiles of DWGS-solutions according to Pedroza, Zalacain, Lara, & Salinas (2010). Free volatiles were determined by immersion of Twister™ (Gerstel, Mülheim an der Ruhr, Germany) polydimethylsiloxane coated stir bar (0.5 mm film thickness, 10 mm length) in 10 mL of each DWGS-solution spiked with 100 μL of internal standard (100μL/L γ-hexalactone and 100 μL/L 3-methyl-1-pentanol). DWGS-solutions were stirred at 500 rpm during 1 h at 25°C. For bound volatile analysis, DWGS-solutions were acidified with 2 M citric acid solution to pH 2.5 prior to introducing the Twister™ stir bar. Following step where samples were spiked with 100 μL of internal standard and stirred at 500 rpm during 2 h at constant temperature of 70±5°C using a water bath.

Stir bars were removed from samples with magnetic tweezers, rinsed with Mili-Q® distilled water, dried with cellulose tissue and finally transferred into thermal desorption tubes for the thermal desorption-GC–MS analysis.

SBSE-thermal desorption-GC-MS analysis

Volatile compounds were desorbed from the stir bar under the following conditions; oven temperature 290°C, desorption time 4 min, cold trap temperature − 30°C and helium inlet flow 45 mL min⁻¹. Volatiles were then transferred into the Hewlett-Packard 6890 (Palo Alto, USA) gas chromatograph coupled to an Hewlett-Packard 3D mass detector (Palo Alto, USA) with a fused silica capillary column SGE BP21 (stationary phase 30 m length, 0.25 mm i.d., and 0.25 μm film thickness) (Ringwood, Australia). The chromatographic programme was set at 40°C (held for 2 min), raised to 230°C at 10°C min⁻¹ and held for 15 min. For mass spectrometry analysis, an electron impact mode (EI) at 70 eV was used. The mass range varied from 35 to 500 u and the detector temperature was 150°C. Identification was carried out using the NIST library and standard spectra. Quantification was based on 5-point calibration curves of respective standards (R ² > 0.95) in a model wine solution previously described. Bound volatile concentration was obtained by subtracting the corresponding free volatile concentration. In order to avoid matrix interferences between the volatile compounds, the MS quantification was carried out in the single ion monitoring (SIM) mode using the characteristic m/z values reported by Zalacain, Marín, Alonso, & Salinas (2007), as well as their quantification curves.

Statistical analysis

All data reported represent the mean of two replicates. SPSS Statistics 17.0 Software (Chicago, USA) was used to evaluate one-way analysis of variance (ANOVA) at p ≤ 0.05. Post hoc Tukey's HSD test was used to distinguish homogeneous subsets of treatments. Canonical discriminant analysis was also used to establish differences between samples and evaluating the importance of different variables on discrimination.

Results and discussions

Waste grape skin dehydration

High moisture content of waste grape skins together with a large production volume are the main factors affecting their exploitation because of spoilage and consequently loss of potential reuse (Cortés et al., 2010). Following results present thermal dehydration as a pre-processing step to enhance handling and avoid spoilage of waste grape skins.

Oven drying treatment at 60°C produced the highest moisture content followed by; control (freeze-drying), 90°C and finally 100°C (Table 1). The moisture content of all samples is similar to other dehydrated foodstuffs considered microbiologically stable. Thermal treatment drastically affected the appearance of samples as those from control and 60°C treatments were more flexible, softer and kept their reddish–purple colour contrary to samples dehydrated at 90°C and 100°C which were darker, stiffer and more fragile. These observations may be attributed in the first place to the higher moisture content of freeze dried and 60°C samples, secondly to structural changes caused by thermal degradation such as porosity and shrinkage. These changes have been associated with hysteresis and in some cases to poor rehydration capacity (Krokida & Philippopoulos, 2005) which may limit the amount of compounds extracted from grape skins.

Table 1. Moisture content (g_water kg⁻¹ _{dry matter}) of waste grape skins according to the dehydration treatment.

Tabla 1. Contenido de humedad (g_agua kg⁻¹ _{base seca}) de hollejos de uva de acuerdo al tratamiento de deshidratación.

Waste grape skin	Raw	Control freeze-drying	Oven drying
			60°C	90°C	100°C
Bobal	1308.79	71.63^b	156.89^a	63.63^b	21.26^c
Garnacha Tintorera	1528.76	88.53^b	145.55^a	67.44^b	16.06^c
Cabernet Sauvignon	1391.77	94.64^b	173.61^a	90.91^b	15.10^c
AMIX	1490.77	94.50^b	119.70^a	81.97^b	18.45^c
Note: Different letters between dehydration columns indicate significant differences (p < 0.05) Nota: Letras diferentes entre los tratamientos de deshidratación indican diferencias significativas (p < 0,05).

After dehydration, dried grape marc was sieved to separate seeds and stalks from grape skins. Sieving allowed us to obtain a considerable amount of dried seeds, stalks and residual grape skins which represented between 35 to 54% of the total dried marc weight. This fact is important towards revaluation of grape marc, because these byproducts have developing applications such as production of grape seed oil, phenolic extracts and grape seed flour (Lutterodt, Slavin, Whent, Turner, & Yu, 2011).

Contribution of DWGS to the chemical composition of a model wine

Maceration of DWGS into synthetic wine solutions allowed extracting colour, phenolic and aroma compounds. After removal of DWGS from the synthetic wine solution, it was observed that 90 and 100°C samples had a residual haze due to low particle size grape skins that could have been formed during maceration. This was associated to the structural damage mentioned previously. Such an effect was observed at a lesser extent in control and 60°C samples. To eliminate haze, DWGS-solutions were centrifuged during 5 min at 4000 rpm. Although residual haze could be considered a problem, common winery operations as settling and simple filtration are enough to remove suspended particles. No further haze was observed in stored samples after centrifugation.

Colour and total polyphenols

The colour intensity of all DWGS-solutions shows the influence of drying temperatures (Table 2). High dehydration temperature treatments and particularly100°C had significantly higher colour intensity values where yellow colour was predominant, indicating a higher phenolic compound degradation. On the other hand, the red component of Cabernet Sauvignon was particularly enhanced by the 60°C treatment although it also experienced the highest degradation when high temperatures were used compared to control. In the case of Garnacha Tintorera and AMIX, no significant differences in the percentage of red colour have been observed between the control and 60°C treatment. Blue colour was randomly affected by dehydration; surprisingly AMIX-60°C solution (the mix of red and white varieties) had the highest blue colour (8.47%) followed by Garnacha Tintorera-Control (7.38%) and Bobal-Control (6.89%). No statistically significant differences have been observed in terms of the blue component between Garnacha Tintorera and Cabernet Sauvignon model wine solutions for any dehydration treatments. Bobal-60 and 90°C solutions had no significant differences against the control.

Table 2. Total polyphenol index (TPI) and colour parameters of dehydrated waste grape skins macerated in synthetic wine solutions.

Tabla 2. Índice de polifenoles totales y parámetros de color de las soluciones de hollejos deshidratados macerados en vino sintético.

		Oven drying
Control	60°C	90°C	100°C
Bobal
TPI	33.71^a,1	25.04^b,3	23.13^c,2	7.21^d,3
Colour intensity	0.01^c,2	0.02^b,2	0.02^b,3	0.03^a,3
Shade	0.64^c,1	0.58^c,3	1.40^b,3	4.51^a,3
% Yellow	36.44^c,1	34.81^c,3	54.42^b,4	79.88^a,3
% Red	56.66^b,1	59.30^a,1	38.68^c,1	17.70^d,2
% Blue	6.89^a,2	5.88^a,1	6.89^a,1	2.41^b,2
Garnacha Tintorera
TPI	23.03^b,c,2	26.55^a,2	23.96^b,1	22.88^c,1
Colour intensity	0.02^d,1	0.03^c,1	0.04^b,1	0.06^a,1
Shade	0.55^c,1	0.55^c,4	1.49^b,3	3.07^a,3
% Yellow	32.87^c,1	32.96^c,4	55.91^b,3	71.63^a,4
% Red	59.90^a,1	59.73^a,1	37.42^b,2	23.26^c,1
% Blue	7.38^a,1,2	7.12^a,1	6.66^a,1	5.10^a,1
Cabernet Sauvignon
TPI	14.78^c,4	43.97^a,1	17.42^b,3	9.05^d,2
Colour intensity	0.02^c,2	0.01^d,3	0.04^a,2	0.03^b,3
Shade	0.68^b,1	0.62^b,2	5.98^a,1	4.71^a,1
% Yellow	36.40^d,1	38.45^c,1	85.67^b,1	97.94^a,1
% Red	53.50^b,1	61.54^a,1	14.32^c,4	2.05^d,4
% Blue	0.01^a,3	0.01^a,2	0.01^a,2	0.08^a,2
AMIX
TPI	19.42^a,3	14.24^b,4	11.20^c,4	0.75^d,4
Colour intensity	0.01^b,3	0.01^b,2	0.02^b,3	0.05^a,2
Shade	0.68^c,1	0.66^c,1	2.40^b,2	8.40^a,2
% Yellow	35.24^c,1	36.41^c,2	65.41^b,2	87.87^a,2
% Red	59.75^a,1	59.10^a,2	27.15^b,3	10.69^b,3
% Blue	6.63^b,2	8.47^a,1	7.43^a,1	1.43^c,2
Note: Different superscript letters within columns indicate significant differences (p < 0.05) between dehydration treatments. For each parameter, different superscript numbers within a row indicate significant differences (p < 0.05) between waste grape skin varieties. TPI is expressed in mg L^−1 of gallic acid equivalents. Control refers to waste grape skins dehydrated by freeze drying. Notas: Letras diferentes entre las columnas indican diferencias significativas (p < 0,05) entre los tratamientos de deshidratación. Para cada parámetro, los superíndices numéricos diferentes entre filas indican diferencias significativas (p < 0,05) entre las variedades de hollejos. TPI está expresado en mg L^−1 de equivalentes de ácido gálico. La liofilización se utilizó como tratamiento control.

Bobal and AMIX DWGS-solutions showed a decrease in the total polyphenol index (TPI) as higher drying temperatures were used (Table 2). In the case of Garnacha Tintorera and Cabernet Sauvignon DWGS-solutions, the highest TPI value was achieved by the 60°C treatment. It was remarkable that Cabernet Sauvignon samples had minor differences of TPI between the control and 90°C treatments and that Garnacha Tintorera-90 and 100°C had no significant differences compared with the control. Dehydration temperature affected the TPI for each variety in a different way as cellular structure/composition might play a fundamental role for protecting phenolic compounds from degradation (Pinelo, Arnous, & Meyer, 2006). Spigno et al. (2007) suggested that the effect of temperature cannot be generalised since it strongly depends on typology of compounds and polymerisation reactions.

Anthocyanins and low molecular weight phenolic compound composition

The anthocyanin profile of Bobal, Cabernet Sauvignon and AMIX wine model solutions showed the typical pattern characterised for having Mv-3-G as the most abundant compound (Cheynier, Moutounet, & Sarni-Manchado, 2003). Garnacha Tintorera DWGS-solutions were distinctive for having the highest concentration of total evaluated anthocyanins as well as containing peonidin-3-glucoside (Pn-3-G) as the most abundant compound, immediately followed by Mv-3-G. Although Garnacha Tintorera skins have been reported to have Mv-3-G as the most abundant anthocyanin closely followed by Pn-3-G, the coloured flesh of this variety is almost entirely dominated by Pn-3-G (Muñoz, González, Alonso, Romero, & Gutiérrez, 2009). Since grape crushing enables contact between pulp and skins, considerable amounts of Pn-3-G might be susceptible to retention in skins during the pressing process thus explaining the slightly higher proportion of Pn-3-G in our samples. The anthocyanin profile of AMIX is quite important due to the red variety contribution (Table 3), which due to the similarities with the other tested DWGS, it may be Bobal variety. The highest concentration of acylated anthocyanins were found in Garnacha Tintorera samples. Cabernet Sauvignon solutions were distinctive for having a higher proportion of acetyl anthocyanins with respect to coumaroyl anthocyanins.

Table 3. Anthocyanin concentration (mg L⁻¹) of dehydrated waste grape skins macerated in synthetic wine solutions.

Tabla 3. Concentración de antocianos (mg L⁻¹) de las soluciones de hollejos deshidratados macerados en vino sintético

		Oven drying
	Control	60°C	90°C	100°C
Bobal
Dephinidinl–3-glucoside	4.46^a,1	1.83^b,1	1.37^c,1	0.12^d,3
Cyanidin-3-glucoside	2.75^a,1	0.71^b,3	0.42^c,1	0.04^d,2
Petunidin-3-glucoside	4.07^a,1	1.40^b,3	0.78^c,1	0.11^d,1
Peonidin-3-glucoside	7.22^a,2	3.50^b,2	1.68^c,2	0.47^d,2
Malvidin-3-glucoside	11.76^a,3	6.90^b,2	3.41^c,2	0.72^d,3
Total non-acyl anthocyanins	30.27^a,2	14.38^b,2	7.68^c,2	1.48^d,3
Total acetyl-anthocyanins	1.80^a,2	0.97^b,3	0.55^b,1	0.19^c,2
Total coumaroyl-anthocyanins	8.19^a,2	3.70^b,2	2.49^c,2	0.72^d,2
Total pigments	40.26^a,2	19.05^b,2	10.72^c,2	2.39^d,2
Garnacha Tintorera
Dephinidinl-3-glucoside	2.51^a,3	1.83^b,1	0.58^c,2	0.33^d,2
Cyanidin-3-glucoside	2.48^a,2	1.66^b,1	0.18^c,2	0.05^d,1
Petunidin-3-glucoside	3.20^a,3	2.39^b,1	0.38^c,2	0.11^d,1
Peonidin-3-glucoside	20.37^a,1	13.87^b,1	3.69^c,1	1.27^d,1
Malvidin-3-glucoside	18.33^a,1	13.14^b,1	3.86^c,1	1.35^d,1
Total non-acyl anthocyanins	46.91^a,1	32.91^b,1	8.71^c,1	3.13^d,1
Total acetyl-anthocyanins	1.67^a,2	1.13^b,3	0.34^c,2	0.13^c,2
Total coumaroyl-anthocyanins	13.33^a,1	8.90^b,1	3.50^c,1	1.37^c,1
Total pigments	61.91^a,1	42.94^b,1	12.54^c,1	4.63^d,1
Cabernet Sauvignon
Dephinidinl-3-glucoside	4.19^a,2	1.71^b,2	0.41^c,3	0.43^c,1
Cyanidin-3-glucoside	1.40^a,4	0.20^b,4	0.03^c,3	0.03^c,3
Petunidin-3-glucoside	3.38^a,2	0.89^b,4	0.15^c,4	0.11^c,1
Peonidin-3-glucoside	3.00^a,4	1.10^b,4	0.21^c,4	0.17^c,3
Malvidin-3-glucoside	11.94^a,2	5.53^b,4	0.97^c,3	0.78^c,2
Total non-acyl anthocyanins	23.93^a,3	9.51^b,4	1.79^c,3	1.54^c,2
Total acetyl-anthocyanins	7.00^a,1	3.48^b,1	0.75^c,1	0.52^c,1
Total coumaroyl-anthocyanins	4.93^a,3	1.71^b,3	0.77^c,3	0.45^c,3
Total pigments	35.86^a,3	14.70^b,3	3.31^c,3	2.51^c,2
AMIX
Dephinidinl-3-glucoside	1.98^a,4	1.65^b,2	0.24^c,4	0.01^d,4
Cyanidin-3-glucoside	1.61^a,3	1.12^b,2	0.17^c,2	0.01^d,4
Petunidin-3-glucoside	1.90^a,4	1.47^b,2	0.21^c,3	0.01^d,2
Peonidin-3-glucoside	4.32^a,3	2.98^b,3	0.52^c,3	0.04^d,4
Malvidin-3-glucoside	8.30^a,4	6.46^b,3	1.07^c,3	0.06^d,4
Total non-acyl anthocyanins	18.13^a,4	13.68^b,3	2.23^c,3	0.1^d,4
Total acetyl-anthocyanins	1.78^a,2	1.49^b,2	0.27^c,2	0.01^d,3
Total coumaroyl-anthocyanins	4.60^a,3	2.51^b,2,3	0.52^c,3	0.01^d,4
Total pigments	24.51^a,4	17.68^b,2	3.02^c,3	0.01^d,3
Notes: Different superscript letters within columns indicate significant differences (p < 0.05) between dehydration treatments. For each compound, different superscript numbers within a row indicate significant differences (p < 0.05) between waste grape skin varieties. Control refers to waste grape skins dehydrated by freeze drying. Notas: Letras diferentes entre las columnas indican diferencias significativas (p < 0,05) entre los tratamientos de deshidratación. Para cada compuesto, los superíndices numéricos diferentes entre filas indican diferencias significativas (p < 0,05) entre las variedades de hollejos. La liofilización se utilizó como tratamiento control.

Anthocyanin concentration was significantly damaged by the dehydration temperatures of the raw material as observed in DWGS-solutions. Piffaut, Kader, Girardin, & Metche (1994) suggested that thermal dehydration at 100°C influence anthocyanin equilibrium by shifting the flavylium form (coloured) towards the chalcone form (non-coloured). These degradation products may be responsible for the lower red component detected by UV-vis analysis in thermal DWGS solutions.

Oven drying at 60°C was the thermal treatment causing the least anthocyanin degradation in terms of total anthocyanins. The influence of the red variety on AMIX solutions was remarkable because it had a total anthocyanin concentration higher than those from Cabernet Sauvignon within the 60°C treatment. Total anthocyanin loss at 60°C in AMIX (29%) and in Garnacha Tintorera (31%) solutions was significantly smaller from those of Bobal (54%) and Cabernet Sauvignon (54%). This could be attributed in Garnacha Tintorera to the higher initial concentration and in AMIX to an improved resistance to degradation due to stabilisation reactions of anthocyanins with other phenolic compounds such as (+)-catechin (Boulton, 2001). Bobal-90°C and Garnacha Tintorera- 90°C solutions had a 73% and 80% loss of total anthocyanins respectively, while all DWGS-100°C solutions lost more than 90%. Similar degradation was observed by Sadilova, Stintzing, & Carle (2006) when applying a 95°C thermal treatment during 4 h to anthocyanin extracts from different foodstuffs.

When comparing 60°C solutions of Bobal and Cabernet Sauvignon with the control, a reduction of 45% for all non-acylated anthocyanins was noted, except for Mv-3-G in Bobal, where a 37% loss was detected. It was observed that Cy-3-G was the most sensitive monoglucoside to oven drying since it experienced the highest loss in all studied DWGS solutions possibly due to its o-dihydroxy structure favouring degradation. In terms of acylated anthocyanins, acetyl derivates were more temperature stable than their respective coumaroyl derivatives. Increasing methoxylation of the acyl moiety of anthocyanins has been associated to structural integrity improvement towards heat treatment (Sadilova et al., 2006).

Low molecular weight phenolic compounds (LMWPC) are closely related to; antioxidant properties, co-pigmentation phenomena, colour stability and organoleptic profile of foodstuffs (Boulton, 2001). (+)-Catechin, (−)-epicatechin, ferulic acid, trans-resveratrol and vanillic acid were the only compounds detected from all samples, although other ones were also sought.

Regarding LMWPC profile, there were significant differences among all DWGS solutions (Table 4). Samples from Garnacha Tintorera and Bobal showed more compounds than the balance of the samples. Remarkable concentrations of (−)-epicatechin and (+)-catechin were found in Garnacha Tintorera solutions, as well as vanillic acid and (+)-catechin in Bobal. Cabernet Sauvignon solutions had the poorest LMWPC composition but had a particularly high concentration of trans-resveratrol. It is known that this compound mainly accumulates in grape skins and has received much attention for its antitumoral and antioxidant bioactivity (Versari, Paola Parpinello, Battista Tornielli, Ferrarini, & Giulivo, 2001; Zhuang, Kim, Koehler, & Doré, 2003).

Table 4. Low molecular weight phenolic compounds concentration (mg L⁻¹) of dehydrated waste grape skins macerated in synthetic wine solutions.

Tabla 4. Concentración de compuestos fenólicos de bajo peso molecular (mg L⁻¹) de las soluciones de hollejos deshidratados macerados en vino sintético.

		Oven drying
	Control	60°C	90°C	100°C
Bobal
(+)-Catechin	1.65^a,2	1.06^b,2		n.d.
Vanillic acid	2.63^1	n.d.		n.d.
(−)-Epicatechin	n.d.	n.d.		n.d.
Ferulic Acid	0.15^a,1	0.15^a,1	0.18^a,1,2	n.d.
(E)-Resveratrol	0.66^a,2	0.30^b,1	0.21^c,1	n.d.
Total LMWPC	5.11^a,2	1.52^b,1	0.4^c,1	n.d.
Garnacha Tintorera
(+)-Catechin	1.83^2	n.d.	n.d.	n.d.
Vanillic acid	n.d.	n.d.	n.d.	n.d.
(−)-Epicatechin	6.81^1	n.d.	n.d.	n.d.
Ferulic acid	0.13^b,1	0.01^b,2	0.21^a,1	0.19^a,1
(E)-Resveratrol	0.04^b,3	0.12^a,2,3	0.06^b,2	0.02^c,1
Total LMWPC	8.8^a,1	0.14^c,2	0.27^b,2	0.21^b,1
Cabernet Sauvignon
(+)-Catechin	n.d.	n.d.	n.d.	n.d.
Vanillic acid	n.d.	n.d.	n.d.	n.d.
(−)-Epicatechin	n.d.	n.d.	n.d.	n.d.
Ferulic acid	0.18^a,1	0.16^a,1	n.d.	0.07^a,b,2
(E)-Resveratrol	2.77^a,1	0.22^b,1,2	n.d.	n.d.
Total LMWPC	2.9^a,3	0.39^b,2	n.d.	0.07^c,2
AMIX
(+)-Catechin	2.12^a,1	1.39^b,1	n.d.	n.d.
Vanillic acid	n.d.	n.d.	n.d.	n.d.
(−)-Epicatechin	n.d.	n.d.		n.d.
Ferulic acid	0.20^a,1	0.20^a,1	0.13^b,2	n.d.
(E)-Resveratrol	0.11^a,3	0.04^b,3	n.d.	n.d.
Total LMWPC	2.44^a,3	1.64^b,1	0.13^7c,3
Notes: Low molecular weight phenolic compounds (LMWPC). Different superscript letters within columns indicate significant differences (p < 0.05) between dehydration treatments. For each compound, different superscript numbers within a row indicate significant differences (p < 0.05) between waste grape skin varieties. Control refers to waste grape skins dehydrated by freeze drying. Notas: Compuestos fenólicos de bajo peso molecular (LMWPC). Letras differentes entre las columnas indican diferencias significativas (p < 0,05) entre los tratamientos de deshidratación. Para cada compuesto, los superíndices numéricos diferentes entre filas indican diferencias significativas (p < 0,05) entre las variedades de hollejos. La liofilización se utilizó como tratamiento control.

Hardly any LMWPC were released into hydroalcoholic solutions when they were prepared with DWGS from high temperatures (90 and 100°C). Nevertheless, ferulic acid was the only compound having ambiguous results since Garnacha Tintorera solutions released the highest concentration at 90°C and 100°C. Trans-resveratrol concentration in solutions was particularly affected as it was found a 92% loss at 60°C and complete degradation at 90 and 100°C. Versari et al. (2001) found that trans-resveratrol degradation in grapes may be accentuated by long term thermal dehydration, reducing temperatures and drying times may preserve this compound.

As the objectives of the present work were focused on major phenolic compounds, other compounds like flavanol glycosides and hydroxycynnamoyltartaric acid derivates were not evaluated. However, presented observations in Table 4 are useful to predict that due to their low concentration and the degradation effect of thermal dehydration, samples might not contain them.

Particular characteristics of DWGS such as; high anthocyanin concentration and profile in Garnacha Tintorera, resveratrol in Cabernet and catechin from Bobal, might be used in mixtures e.g. AMIX, for producing tailor made products aiming to improve particular sensory and nutraceutical properties of beverages and other foodstuffs.

Volatile composition

The volatiles released by DWGS into model wine solutions were detected in very low concentration values (Figure 1). β-ionone and β-damascenone (norisoprenoid) with floral descriptor; D-limonene, α-terpineol and geranyl acetone (terpenes) with citrus, floral and fruity notes; and (E)-2-hexenal (C6 volatiles) with a herbaceous note were detected in DWGS-solutions (Figure additional) although other volatiles were also evaluated. These compounds have been regarded as characteristic odorants of red wines made with similar varieties (López, Ferreira, Hernández, & Cacho, 1999).

Figure 1. Volatile composition and standard deviation after oven dehydration from different dehydrated waste grape skin model wine solutions.

Figura 1. Composición volátil y desviación estándar de las soluciones de hollejos deshidratados macerados en vino sintético.

Dehydration temperature significantly affected the volatile compounds released in DWGS-solutions with the 60°C treatment producing a release of terpenes in Bobal, Cabernet and AMIX samples; this group of compounds was also the most important in terms of concentration among these samples (Figure 1). Control (freeze-drying) solutions of all DWGS were distinguished by the presence of (E)-2-hexenal although the range of concentration was not over its olfactory threshold (0.6 mg L⁻¹) in any of thermal treatment samples (Etiévant, 1991). Control treatment of Garnacha Tintorera revealed a concentration of β-ionone (0.31 μg L⁻¹) more than three times over the olfactory threshold. This compound remained without significant differences after all dehydration treatments of Garnacha Tintorera. The 60°C samples of Bobal, AMIX and Cabernet Sauvignon had a higher concentration of terpenes while Garnacha Tintorera lost all volatiles but β-ionone. D-limonene was found in all DWGS-solutions of 90°C and 100°C, excluding Garnacha Tintorera.

Wine solution from Cabernet Sauvignon DWGS-60°C released a significant amount of β-ionone (0.32 μg L⁻¹) after thermal treatment. As exposed by Crouzet, Kanasawud & Sakho (2002), thermal treatment induces carotenoid degradation products such as β-ionone. This observation is particularly important because it implies that the optimisation of time/temperature parameters may favour the generation of these low odour-threshold impact odorants. Furthermore, the release of α-terpineol in Bobal-60°C samples (7.06 μg L⁻¹) might as well be related to thermal dehydration, α-terpineol has been reported to be the final degradation/stabilisation product of different terpenols (Williams, Strauss, Wilson, & Massy-Westropp, 1982). Although terpenols released into solutions are not over the olfactory threshold, they may contribute to improve the aroma perception of fruity, citrus and floral aromatic series.

Bound aroma fraction evaluated in all solutions was limited quantitatively and qualitatively. No bound odorants have been detected in Cabernet Sauvignon or Garnacha Tintorera DWGS solutions. β-ionone was found in all Bobal solutions (0.05 μg L⁻¹), without significant differences between control, 60°C and 90°C (data not shown). Only AMIX-control solutions reported a bound fraction composition of trans-2-hexenal (57.85 μg L⁻¹), β-ionone (0.05 μg L⁻¹), and β-damascenone (0.18 μg L⁻¹).

The low concentration of free and bound volatiles might be related to the “non-aromatic” nature of grape skin varieties (Bayonove, Baumes, Crouzet, & Günata, 2003), as well as to the preliminary extraction of these compounds at juice factory and volatilisation/degradation during dehydration.

Canonical discriminant analysis was performed using all measured parameters in order to evaluate differences and similarities between dehydration treatments and varieties. Figure 2a shows the differences among dehydration treatments, where 99% of variance was described by two canonical functions (CF) (CF1 = 83.9% CF2 = 15.1%). As the most important function coefficients in both canonical functions were anthocyanins, a good separation was achieved among treatments regardless of the variety. This can also be observed through the homogeneity of the groups, indicating independently of the variety that the overall effect of drying was similar. These figures also confirm previous observations on the similarities among DWGS solutions from high temperature treatments (90 and 100°C) and the intermediate characteristics of DWGS-60°C with respect to the control dehydration by freeze-drying.

Figure 2. Discriminant analysis scatter plots of dehydration treatments and dehydrated waste grape skins (DWGS).

Figura 2. Análisis discriminante de las soluciones de hollejos (DWGS) de acuerdo al tratamiento de deshidratación y a la variedad.

Dispersion of points within varieties (Figure 2b) shows the influence of dehydration treatments, nevertheless each DWGS can be distinguished. The 98% of variance was described by two CFs where anthocyanins had the most important canonical discriminant coefficients. Garnacha Tintorera solutions were separated from the rest of the samples, while Cabernet Sauvignon, Bobal and AMIX were closer. Discrimination of varieties could then be associated to the particular profile characteristics like the high content of peonidin-3-glucoside in Garnacha Tintorera DWGS.

Direct addition of DWGS to white wine

An assay using a commercial white wine of Airén variety was done to evaluate the release of compounds from DWGS in a real sample. DWGS of Bobal-60°C were selected for having a high concentration of LMWPC and volatile compounds. The maceration of DWGS with white wine produced rosé wine with a colour intensity (0.02) and shade (0.61) similar to those observed in synthetic wine samples. Figure 3 shows a comparison among Bobal-60°C in synthetic wine, white wine (without DWGS) and rosé wine produced after direct addition of Bobal-60°C. A higher yield of TPI was detected in Bobal-60°C rosé wine than that obtained with synthetic wine, indicating that other polyphenols are susceptible to extraction. Important bioactive compounds such as catechin and resveratrol were released to the same extent as in synthetic wines, contributing to a controlled extraction of bioactive compounds. The volatile composition was not considerably improved (Figure 3), although β-ionone and geranyl acetone were released by Bobal DWGS. A loss of D-limonene and β-damascenone was noticed in DWGS-rosé wine indicating that degradation or volatilisation occurred during the process. Other parameters related to mass transfer phenomena such as particle size, dosage, maceration time, etc. are out of the scope of the present work and will be researched in further studies.

Figure 3. Composition and standard deviation of synthetic and rosé wine prepared with dehydrated waste grape skins (DWGS) of Bobal.

Figura 3. Composición y desviación estándar de los vinos rosados elaborados con hollejos deshidratados (DWGS) de Bobal.

Supplementary Figure 1. Free aroma fraction of hydroalcoholic solution after direct addition of dehydrated waste grape skins.

Figura adicional 1. Fracción libre de los aromas presentes en las soluciones de hollejos (DWGS) de acuerdo al tratamiento de deshidratación y a la variedad.

Conclusions

Thermal dehydration of juice waste grape skins prevented spoilage, enhanced handling and storage leading to efficient exploitation. Oven dehydration caused the loss of colour, phenolic and aroma compounds from such waste grape skins. Regardless of variety, dehydrated samples at 60°C had the highest red colour and released the highest amount of anthocyanins. Low molecular weight phenolic compounds were particularly affected by dehydration as almost complete degradation was observed in all DWGS. Thermal dehydration caused the release of important odorants like β-ionone and α-terpineol into solutions which can contribute to the volatile composition of wines. Garnacha Tintorera was the variety with the highest amount of anthocyanins, while Bobal had the highest amount of resveratrol after thermal dehydration. Bobal-60°C and AMIX-60°C samples were more resistant to thermal degradation of phenolic compounds than the balance of samples.

The assay with white wine revealed that DWGS can be used as a direct additive to produce rosé wines, allowing a controlled extraction of compounds and an innovative/sustainable reuse of waste material from the grape industry. Optimisation of extraction conditions and other applications for the food industry deserve further research.

Supplementary material

The supplementary material for this article is available online at http://dx.doi.org/10.1080/19476337.2011.633243.

Acknowledgements

This work was financially supported by the Junta de Comunidades de Castilla-La Mancha (Project PAI08-0148-9842). M.A. Pedroza has received a grant from Consejo Nacional de Ciencia y Tecnología of the Mexican Government. Thanks to Ana Soler at Julian Soler S.A. Juice Concentrate factory (Quintanar del Rey, Cuenca, Spain) for her technical assistance and for supplying waste grape skins. Thanks to Kathy Walsh for proof-reading the English manuscript.